The AIDS epidemic is a world wide problem. Over the past three decades, HIV infections have risen from sporadic cases to a global pandemic involving more than 18 million individuals. In 1995 alone, end stage infections (AIDS) killed one million individuals, making it the world's sixth largest leading cause of death by infectious disease. Even worse, all indicators suggest that AIDS deaths will continue to climb in this ranking. An estimated seventy-five percent of the world's new infections are attributable to sexual transmission, making it imperative that we understand the virologic factors influencing the pandemic.
Investigations have shown that HIV mutates very rapidly--as much as 10 to 50 times faster than an influenza virus. Further, some of the most virulent strains of HIV are appearing in remote corners of the globe. For example, HIV-1 strains have been observed in Thailand which appear to be more infectious than the more common HIV-1 strain prevalent in the western world.
There are no vaccines to prevent the spread of an HIV infection and their timetable for development is uncertain. Antiviral drugs (such as AZT and the latest protease inhibitors) may help to prolong the lives of infected individuals (or increase their quality of life) but they have no impact on preventing new infections. Much progress has been made in identifying host HIV infection spreads (sexual intercourse, mother to child, needle sharing by drug users, and blood transfusions) and in launching public health programs to reduce these modes of transmission. But the important strides made thus far in controlling the epidemic may ultimately be limited by HIV's apparent ability to mutate rapidly and to increase its transmissibility.
The spread of HIV depends on the behaviors and interactions of people as well has the inherent transmissibility of the virus. In developed countries, epidemiological studies show that the risk of person-to-person transmission ranges from one to five new infections per 1000 sexual encounters. In developing countries like Thailand, more recent epidemiological studies suggest that the risk of transmission is one order of magnitude greater--ten to fifty new infections per 1000 encounters. Attempts to attribute this increased transmission to known HIV risk factors such as the numbers of sexual partners, frequency of encounters, varieties of behavior, amounts of drug use, and prevalence of sexually transmitted diseases have revealed no clear connection. The world is thus confronted with the worrisome possibility that certain HIV isolates are appearing which are more transmittable than others. This is accompanied by reports of a rapidly growing HIV epidemic in Thailand, which now appears to involve viral isolates that belong to particular genetic subtypes.
Thus far, research on HIV has focused on its genetic and immunologic properties. The RNA genome within HIV mutates very rapidly, primarily due to the error-prone activity of reverse transcriptase. This enzyme produces nearly one base substitution per 3000 nucleotides, which means that each newly transcribed virus contains several mutations in its 10,000-base genome. The high mutation rate has produced countless numbers of HIV variants and, as time passes, it is feasible that more transmissible viruses may appear in concert with accelerating epidemics. This may help to explain the explosive HIV epidemic in Thailand.
Studying the RNA message within HIV has resulted in the cataloguing of thousands of HIV sequences in the Human Retrovirus and AIDS Database. At present, there are at least eight sequence subtypes for HIV-1 (designated by the letters A, B, C and so forth) and, given more limited data, there appear to be at least two for HIV-2. Each subtype has a characteristic phylogenetic map and differing geographic distribution, making it possible to track the evolution of the epidemic. The total number of sequences that may fit into a particular subtype, however is truly enormous. For example, assuming that HIV is limited to utilizing just 2 different amino acids at certain positions within its proteins and that a "model" subtype is determined by substitutions in 30 independent positions (or approximately 1% of the 3000 amino acids in HIV's entire genome), the model subtype could contain as many as 2.sup.30 =10.sup.9 different sequences--an insurmountable number even if only a small set of all possible combinations produced active viruses. Given the difficulty of sequencing thousands of HIV isolates in their entirety, it seems improbable that current technologies will enable identification of genetic sequences that correlate reliably with transmissibility. In other words, certain subtypes may act as surrogate markers of transmissibility but may not identify the underlying mechanisms.
Serotyping HIV isolates is being carried out currently with standardized panels of immunoglobulins that block the virus from infecting of CD4.sup.+ cells. This often used approach has allowed virologists to categorize isolates according to their patterns of susceptibility and to target certain HIV serotypes for vaccine development. Susceptibility to blocking, however, varies markedly with (even) single amino acid substitutions within the gp120 envelope glycoprotein. Consequently, the enormous combinatorics and limitations on sampling has restricted our ability to identify serotypes that correlate reliably with transmissibility.
Epidemiologists are using increasingly convenient tools to conduct rapid field surveys of HIV prevalence and the incidence of new infections. Newer methods use body fluids such as saliva (instead of blood) to determine whether individuals are HIV infected. Because this reduces the physical invasiveness of sample collection, it often increases the willingness of individuals to participate in epidemiological surveys. These innovations, together with various types of mathematical models, permit epidemiologists to track and analyze the HIV epidemic with increasing accuracy. The necessary tools are available for conducting repeated estimates of person-to-person transmissibility and matching them with the physical properties of HIV isolates.
As the above demonstrates, there is a need for new measurement-based schemes that integrate epidemiologic, immunologic and virologic data to understand why more transmissible HIV isolates may be emerging. Unfortunately, this need is difficult to meet for two reasons. First, many of the newer strains of HIV and other infectious diseases are emerging in remote areas of the world where laboratory facilities are simply non-existent. Second, many of the tests which need to be performed to gather the necessary data require an overwhelming number of repetitive operations and a biohazard safe environment.
Current automated laboratory instruments have not been up to the task. These special purpose, automated laboratory instruments have been designed to imitate human actions as closely as possible. A control program directed robotic hands on tracks to shuttle samples from one location to another, and also commanded the same hands to carry out elementary operations. The number of operations determined the number of stations, and procedures were completed when samples reached the end of the machine. These forerunners of today's automated instruments were simple in concept but they also had drawbacks. Their mechanisms were easily overloaded by too many elementary operations and, due to this restriction, overall utility was rather limited. Furthermore, small mechanical glitches (such as the mishandling of samples) often precipitated crashes, necessitating time-consuming interventions by trained technicians.
In recent years, stations performing only one elementary operation have given way to "standard laboratory modules" (SLMs) that integrate several operations into one logical and coherent task. This newer type of design has noteworthy advantages. For example, it distributes work that is ongoing within the instrument among several semi-autonomous units. Placing two or more identical SLMs within the same instrument is analogous to adding extra processor chips to parallel-processing computers--it eliminates bottlenecks and speeds up critical tasks. Decentralization of tasks in SLM-based instruments reduces the number of operations performed by robotic hands, allowing them to act mainly as transporters which markedly increasing instrument flexibility and reliability. SLMs are more easily reconfigured to perform a wider variety of tasks to a wider variety of potential users than special purpose robotics.
However, SLM technology alone is insufficient to meet emerging data acquisition and analysis needs. While SLMs can be applied to broader ranges of experiments, their flexibility remains limited. Further, SLM technology is available only to a small number of research scientists, and will likely remain so because of cost and logistical constraints. In light of the above, it is apparent that there is a need for an integrated SLM technology that provides a broad range of services to the scientific community, especially those adaptable to solve rapidly evolving test analysis problems.